Re: [PATCH v3] x86: Remove libopcodes dependency

[Jan Beulich <jbeulich@suse.com>]

On 02.12.2022 18:19, H.J. Lu wrote:
> On Thu, Dec 1, 2022 at 11:16 PM Jan Beulich <jbeulich@suse.com> wrote:
>>
>> On 01.12.2022 19:26, H.J. Lu wrote:
>>> On Thu, Dec 01, 2022 at 08:41:28AM +0100, Jan Beulich wrote:
>>>> I understand you prefer it, but given you've found what's wrong
>>>> yourself, I'm puzzled by "I don't see anything wrong with my
>>>> approach". Or wait - looks like I misread what you pointed out
>>>> above: That looks to be something from the top level Makefile.
>>>> Yet then again I can't spot any such in my build trees. Where is
>>>> that excerpt from? (I can spot somewhat similar patterns, but
>>>> they're used strictly to recurse _down_, just like looks to be
>>>> the case with what you've quoted.)
>>>>
>>>> In any event, a practical manifestation of the issue I've said
>>>> I'm concerned of is this: If a 2nd party anywhere in the tree did
>>>> the same you do, and if the two $(MAKE) invocations then raced
>>>> with one another, then it's plain undefined what would happen to
>>>
>>> Fixed.
>>
>> In which way? You still ...
>>
>>> --- a/gas/Makefile.am
>>> +++ b/gas/Makefile.am
>>> @@ -448,6 +448,18 @@ development.exp: $(BFDDIR)/development.sh
>>>       $(EGREP) "(development|experimental)=" $(BFDDIR)/development.sh  \
>>>         | $(AWK) -F= '{ print "set " $$1 " " $$2 }' > $@
>>>
>>> +config/tc-i386.@OBJEXT@: $(srcdir)/../opcodes/i386-init.h \
>>> +     $(srcdir)/../opcodes/i386-tbl.h
>>> +
>>> +# i386-gen will generate both headers in one go.  Use a pattern rule to
>>> +# properly express this, with the inner dash ('-') arbitrarily chosen to
>>> +# be the stem.
>>> +$(srcdir)/../opcodes/i386%init.h $(srcdir)/../opcodes/i386%tbl.h: @MAINT@ $(srcdir)/../opcodes/i386-opc.tbl \
>>> +     $(srcdir)/../opcodes/i386-reg.tbl \
>>> +     $(srcdir)/../opcodes/i386-opc.h \
>>> +     $(srcdir)/../opcodes/i386-gen.c
>>> +     $(MAKE) -C ../opcodes gen-i386-tbl
>>
>> ... wrongly recurse from gas/ into opcodes/. In fact I can't spot any
>> change in regard to the cross-subdir operation that I continue to
>> object to. I'm afraid I'm unaware of ways to address this other than
>> by avoiding the recursive $(MAKE) invocation altogether.
> 
> $(srcdir)/../opcodes/i386%init.h $(srcdir)/../opcodes/i386%tbl.h
> will invoke $(MAKE) only once.

But that was a separate problem I was complaining about. This does in no
way address the issue of inappropriate make recursion.

Unless I see a technically correct version from you by the end of this
week, I will commit my variant (series) early next week. It is pretty
unacceptable for you to block my (also further) work by insisting on the
use of a broken alternative patch. I've clearly indicated that I'm okay
with any solution that's better than mine, but such a solution has to be
free of (pretty obvious) flaws.

Jan